Pattern discovery in UTM library circulation database

Abstract

Huge databases are being used in organizations to store data. These databases contain hidden patterns which can be discovered and used in the organizations. In this project, we applied data mining techniques to uncover the patterns in the circulation database of UTM library. In order to discover worthwhile patterns we followed knowledge discovery process (KDD) to transform row data to suitable format. Weka machine learning software was applied to do the data mining task. In this project, we studied two association rules mining algorithms, Apriori and FPGrowth. The later was used to discover some patterns among borrowed books. These patterns which are presented in a list can be used to make recommendations to patrons who are searching for a certain topic based on items that previously were borrowed together. In addition, a novel rule matrix was presented to store the found rules for future use. Both the list for recommendation and rule matrix are useful to construct a recommender system for users of UTM library.